skills.rest
matt-metivier
matt-metivierCOMMUNITY
Explore all skills by this author →
GitHub

Agent Team Spawner

Decompose complex tasks into parallel subtasks and dispatch them across agents.

Updated Aug 27, 2026
Search Tags:#workflow#multi-agent#worktree#parallel#task-decomposition#nats#agent-allocation
One-click install
npx skills add https://github.com/matt-metivier/zk-hub --skill agent-team-spawner
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: Agent Team Spawner
Source: https://github.com/matt-metivier/zk-hub/tree/main/skills/orchestrator/agent-team-spawner
Command: npx skills add https://github.com/matt-metivier/zk-hub --skill agent-team-spawner

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Guides the hub orchestrator on when and how to parallelize work across multiple agents for complex tasks.

Core Features & Use Cases

  • Decompose complex tasks into parallelizable subtasks
  • Allocate subtasks across agents based on availability and affinity
  • Ensure per-subtask worktrees and isolation to prevent conflicts
  • Aggregate evidence and results for informed decision making

Quick Start

Start a parent task and let the spawner automatically decompose and dispatch subtasks to multiple agents.

Frequently Asked Questions about Agent Team Spawner

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I parallelize complex software tasks across multiple agents?▼

Parallelizing complex tasks involves decomposing them into independent subtasks, allocating them across agents based on availability, and enforcing per-subtask worktrees for isolation.

How does load-aware agent allocation work for parallel workflows?▼

Load-aware agent allocation distributes subtasks by matching them to available agents with appropriate capacity and affinity, accelerating delivery through efficient parallel execution.

Do I need NATS to dispatch subtasks to multiple agents?▼

NATS is used as the dispatch mechanism to route subtasks to agents, while SQLite and Temporal back evidence aggregation to ensure end-to-end traceability across the workflow.

When should I use multi-agent task decomposition for feature development?▼

Multi-agent task decomposition is ideal for workflows with high decomposability and several independent subtasks, such as feature development, testing, or data processing across a fleet.

What is the best way to isolate parallel agents working in the same repository?▼

The best way to isolate parallel agents is by enforcing per-subtask worktrees, which prevents conflicts and ensures each agent operates independently within the shared repository.

How is evidence aggregated from parallel agent executions?▼

Evidence and results from parallel agent executions are aggregated using SQLite and Temporal-backed systems, providing informed decision making and end-to-end traceability for the parent task.

Related Skills

Explore More Software Engineering Skills→
drdave-flexnetosdrdave-flexnetos

Messaging & Event Systems

Implement NATS pub/sub and Temporal workflows for event-driven architectures.

Community
Intermediate
bengousbengous
3

layer-testing

Generate layer-focused tests with coverage analysis and isolated git worktrees.

Community
Advanced
vuralserhat86vuralserhat86
43

message_queues

Implements asynchronous communication patterns using Kafka, RabbitMQ, NATS, Redis Streams, Celery, BullMQ, Temporal, and more.

Community
Advanced
Fancu1Fancu1

dispatching-parallel-agents

Dispatch independent agents to run concurrently on distinct problems.

Community
Intermediate